Graph-powered link analysis
Traditional link analysis relies on table-based data models, where relationships are derived through joins. As networks grow more complex, this approach can become computationally expensive and makes multi-hop connections harder to explore.
Graph-based link analysis stores relationships as first-class elements of the data model. This enables efficient multi-hop traversal, more flexible schema evolution, and easier integration of additional data sources.
In real-world investigations, link analysis often involves gigabytes of interconnected data. Traversing even a few ownership layers using repeated table joins across large datasets can become memory-intensive and computationally expensive.
GraphAware’s link analysis software uses native graph queries to traverse relationships directly, enabling efficient multi-step exploration without reconstructing the network. Because relationships are stored explicitly, performance is driven by the connectivity of the data rather than the total dataset size, supporting consistent traversal even as network depth increases.
Lack of persisted relationships
In investigative contexts, link analysis frequently requires answering time-based questions — for example, who owned an asset at a specific point in time. This depends on capturing when relationships start and end.
In relational systems, relationships are not stored as first-class entities, so answering new temporal questions often requires additional joins and reconstruction of ownership chains.
GraphAware’s link analysis software stores both nodes and relationships with their own properties, including start and end dates. Adding a temporal dimension typically involves applying a time-based condition to the query, rather than rebuilding the structure.
The graph model also supports flexible exploration. New data sources — such as transactions or regulatory filings — can be incorporated without restructuring the dataset, enabling ongoing discovery of emerging patterns and connections.
